Meta hires and fires AI safety team in just 4 months: What happened?

Meta is letting go of employees it hired from AI security startup Virtue AI just four months after bringing them into its Superintelligence Labs. The decision comes after Meta had brought in the team to improve its work on AI safety and security.Meta spokesperson Andy Stone told Semafor that the arrangement did not work out as expected because of differences in work styles."Unfortunately, the arrangement didn’t work out as planned," Stone said. He added that Meta Superintelligence Labs continues to work on AI safety, alignment and frontier risk.Meta hired Virtue AI's founders in JuneIn June, Meta hired Virtue AI co-founders Bo Li, Dawn Song and Sanmi Koyejo, along with other members of the startup's team. Axios had reported at the time that the group was joining Meta Superintelligence Labs as the company expanded its AI security efforts. Virtue AI had developed security tools for enterprise AI systems, including automated red teaming, runtime guardrails and AI governance. The startup had also previously worked with companies including Anthropic and OpenAI, as well as the US Commerce Department's National Institute of Standards and Technology (NIST).Meta said at the time that keeping its increasingly capable AI systems safe and reliable was a key priority. Li and Song were expected to report to Nat Friedman, an executive at Meta Superintelligence Labs, while Koyejo was to report to Rob Fergus, who heads Meta's fundamental AI research lab, FAIR. The hiring came when AI companies were facing growing questions about the safety of AI agents and more advanced models. Companies were investing more in security as governments in the US and elsewhere paid closer attention to the risks associated with increasingly capable AI systems. Four months later, however, Meta has decided to let the Virtue AI employees go.Meta is also cutting jobs in other segmentsThe decision comes as Meta continues to make large workforce cuts while investing heavily in AI. In April, the company told employees that it planned to eliminate about 10 per cent of its workforce, or roughly 8,000 jobs. Employees were told they would find out on May 20 whether they were affected. The waiting period was difficult for some workers, with former Meta engineering manager Robert Sweeney describing the weeks before the layoffs as a “death march." He resigned a day after the cuts.Meta has also hired groups of employees from other AI startups as it builds its AI teams. In March, for example, the company brought in the founders and team behind Dreamer, an AI startup focused on helping people build AI agents.The Virtue AI case is different because the team joined Meta only in June and is now leaving just four months later. Meta has not provided details about which employees are leaving or whether all former Virtue AI staff members are affected.The Wall Street Journal recently reported that Meta has plans to invest as much as $135 billion in AI infrastructure this year as the company works towards its goal of developing what it calls "personal superintelligence" for its 3.5 billion daily users. This is rumoured to be one of the reasons behind a series of layoffs at Meta.- Ends